Hey，TextResult has a big trouble! about jodd HOT 11 CLOSED

Test Code:

response.setContentType(MimeTypes.MIME_TEXT_PLAIN);
//Haha, characterEncoding is back
response.setCharacterEncoding(madvocConfig.getEncoding());

it works! 😄

from jodd.

Perfect! I always hate when setContentType change encoding :)

I will try to build a testcase on this, so we can have test infrastructure for similar cases for Madvoc.

Thank you! I will soon release a beta snapshot!

from jodd.

Thanx again, i solved it in the same way and also by using OutputStream as I wanted to set the content length as well, in case when server does not provide chunked transfer.

from jodd.

but what about other Result(s)?
how about
MadvocResponseWrapper.setContentType(..)

if (contentTypeResolver.getEncoding() != null){
    characterEncoding = contentTypeResolver.getEncoding(); 
}

from jodd.

In short, the approach is the following: 1) Madvoc sets the content type and encoding first 2) if the Result wants to modify it, than its his responsibility to do everything right.

I know that MadvocResponseWrapper seems like a single place for fixing this; but that would always append the character encoding to content type; and the purpose of the wrapper is to allow to reset it:) Let me explain. In cases when you return raw types, like for images created dynamically, returned content type would always have character encoding set; and it would look like:

content-type:image/jpeg;charset:UTF-8`

Some browsers (like IE) do not like this, and, being strict, that would not be correct. So we introduced MadvocResponseWrapper to allow raw images to be able to reset charset and allow sending of just:

content-type:image/jpeg

See: https://github.com/oblac/jodd/blob/master/jodd-madvoc/src/main/java/jodd/madvoc/result/RawResult.java#L47

Since no other result is changing the content type, I believe we are fine here. Make sense :)?

from jodd.

Well, my JsonResult TemplateResult (instead of jsp) even CachingPageInterceptor have the same problem , I may fix them.
btw, do you think about a CachingPageInterceptor for Jodd Madvoc?

from jodd.

Yes, you are probably using Writer to output the content; while thats correct, try using OutputStream in your results as it would force to use the encoding...

This might be annoying, I agree, but I am not sure how to make a difference when character encoding is missing (i.e. being null) on purpose - like in RawResult for binary types. Let me think a bit more about this :)

Sure, I am thinking about caching, for now the only thing I am not sure is how to implement the cache validity check in most pragmatic way? You can send me your idea, no problem with that :)

from jodd.

it's a piece of cake for me figure out where should reset characterEncoding, however , we should line it out in doc for other guys.

[how about:]
MadvocResponseWrapper override .getCharacterEncoding()
so we can do it like is

String encoding = response.getCharacterEncoding();
response.setContentType(MimeTypes.MIME_TEXT_PLAIN);
response.setCharacterEncoding(encoding);

no

@In(scope = ScopeType.CONTEXT)
MadvocConfig madvocConfig;

from jodd.

Nice 👍 Please let me check first (write a test) for a RawResult and character encodings, to see if this will cover the other use case that actually made me creating MadvocResponseWrapper.

from jodd.

See: https://github.com/oblac/jodd/blob/master/jodd-madvoc/src/main/java/jodd/madvoc/component/MadvocController.java#L105

and #L106

a litte mistake , use encoding instead of madvocConfig.getEncoding()

from jodd.

Haaa, thank you:exclamation: I've just fixed and push the changes - I hate such mistakes, please let me know if you find anything similar :smile:

from jodd.